Closed Bug 1086571 Opened 10 years ago Closed 10 years ago

Investigate and fix test_call_log_all_calls intermittent failures


(Firefox OS Graveyard :: Gaia::UI Tests, defect)

Gonk (Firefox OS)
Not set


(Not tracked)



(Reporter: RobertC, Assigned: RobertC)




(2 files)

test_call_log_all_calls is failing intermittently on Jenkins and on local device for different reasons.
The test fails intermittently on both v2.2 and v2.1
This patch is for the following error:
Traceback (most recent call last):
File "/var/jenkins/1/workspace/flame-kk-319.mozilla-central.ui.functional.smoke/.env/local/lib/python2.7/site-packages/marionette_client-0.8.4-py2.7.egg/marionette/", line 264, in run
File "/var/jenkins/1/workspace/flame-kk-319.mozilla-central.ui.functional.smoke/tests/python/gaia-ui-tests/gaiatest/tests/functional/dialer/", line 64, in test_call_log_all_calls
self.assertEqual(len(call_list), 2)
AssertionError: 1 != 2
Attachment #8511047 - Flags: review?(viorela.ioia)
Attachment #8511047 - Flags: review?(jlorenzo)
Attachment #8511047 - Flags: review?(florin.strugariu)
test_call_log_all_calls started failing recently with the errors present in the following adhoc run:

The first digit of the plivo phone number is missing in the call log. Where there any changes done recently in the dialer app which might cause this?

We could not reproduce the issue locally.
Flags: needinfo?(tchung)
Flags: needinfo?(gsvelto)
I'm not aware of changes to how we handle the call log that might do this. needinfo'ing :fcampo because he might know more.
Flags: needinfo?(gsvelto) → needinfo?(fernando.campo)
See Also: → 1088756
I found the change we've seen since October 20th, it's the test itself. It has been modified to check incoming call and it was landed at that date (see bug 1075547 comment 3).

Here's the reason why we don't reproduce locally. In the US, carriers don't add a + for domestic calls, not like in Romania, Canada or France. Then, some providers, like AT&T, don't add the country code (1, in our test case) but some do, like T-Mobile. That's also the reason why this test was intermittent.

Hence, we shouldn't check for the country code in our test for now. Also, to explain that, let's put a comment in the code, about country codes and US carriers.

I opened bug 1088756 to ask UX what we should do with country codes in the US.
Blocks: 1075547
Flags: needinfo?(tchung)
Flags: needinfo?(fernando.campo)
Comment on attachment 8511047 [details] [review]
Link to Github pull-request:

LGTM. Let's just check for the phone number minus the country code, and the patch will be okay to me.
Attachment #8511047 - Flags: review?(jlorenzo) → review-
Comment on attachment 8511047 [details] [review]
Link to Github pull-request:

Removed country code and started an adhoc run:
Attachment #8511047 - Flags: review- → review?(jlorenzo)
Comment on attachment 8511047 [details] [review]
Link to Github pull-request:

Looks good, r+
Attachment #8511047 - Flags: review?(viorela.ioia) → review+
Comment on attachment 8511047 [details] [review]
Link to Github pull-request:

Tested locally. As a country code can be 3 or 4 digits long, this line ( might be problematic if another SIM card from another country does not display the country code. But I hope bug 1088756 will fix this issue in the meantime.

Attachment #8511047 - Flags: review?(jlorenzo) → review+
Attached file uplift to v2.1
Attachment #8512540 - Flags: review?(viorela.ioia)
Attachment #8511047 - Attachment description: Link to Github pull-request: → Link to Github pull-request:
Attachment #8511047 - Flags: review?(florin.strugariu)
Closed: 10 years ago
Resolution: --- → FIXED
QA Whiteboard: [fxosqa-auto-s2]
You need to log in before you can comment on or make changes to this bug.